Facebook has awarded Shion Guha, an assistant professor at Marquette University, a grant to support his research aimed at understanding and mitigating biases in Wisconsin’s foster care system.
He is one of three recipients to win this year’s international Facebook Computational Social Science award.
Guha’s work is centered on using computational social science methodologies to address the issue of instability in foster care placement in Wisconsin.
The state has made a host of child welfare reform efforts over the past 25 years, stemming from a settlement agreement of a 1993 federal class action lawsuit that alleged mismanagement and widespread disparities in Milwaukee’s child welfare system.
Among the conditions included in the 2002 settlement agreement is a mandate that the state raise its foster care placement stability rate to 90 percent. It hasn’t yet reached that benchmark.
Guha was asked about using his research approach to help address the issue and has spent the past year researching the state’s foster care system in collaboration with SaintA, a provider of foster care and adoption services.
“Over the past year, my research has consisted of doing a lot of background reading and talking to people in the foster care system to understand what is going on,” he said. “We have been trying to understand and study case workers. They are the folks who actually interact with the children; they are the those who are developing the good practices.”
The Facebook grant, which will fund a Ph.D. student researcher position for a year, will allow Guha to move on to the next phase of the project.
“Now I want to develop methodology to see what we can do to at least partially find some solutions,” Guha said.
In addition to using the large database of state data, he is particularly interested in gaining access to a potentially valuable new data set: case workers’ notes.
“Why are some case workers more successful than others?” Guha said. “If you interview case workers, it can be so intangible that they will be unable to articulate why some are better than others. However, case workers actually write very detailed notes and there are apps that they use where they have a huge catalog of these notes, which are timestamped and discuss how they resolved cases and how they reached a positive outcome for the child.”
While computational algorithms historically have been better at dealing with quantitative data than understanding relationships between texts, Guha said his background is in finding “latent or hidden patterns from texts that establish relationships between things and people.”
“There is a lot of research in the foster care system; we aren’t the first to look at the system,” he said. “But we’re probably the first group to look at hidden patterns that exist in the actual notes of case workers. This grant will help develop a methodology that can actually infer those latent patterns from the notes.”
Facebook will not provide data to award recipients or be involved in the research process.
Other researchers to receive grants include:
- Detecting Fake News on Social Networks: Michael Bronstein, Universita della Svizzera italiana
- Validating Theory for Quantifying Peer Influence in Observational Data: Edward McFowland III and Gordon Burtch, University of Minnesota-Twin Cities.